• AMPK;
  • memory CD8 T cell;
  • metabolism;
  • mTOR


  1. Top of page
  2. Abstract
  3. Conflict of interest
  4. References

Adenosine monophosphate-activated protein kinase (AMPK) is a serine/threonine kinase and is crucial for cellular energy homeostasis. The exact role of AMPK during memory CD8+ T-cell differentiation, a process that changes from the metabolically active state of effector T cells to one of quiescence in memory cells is not well understood; however, a report by Cantrell and colleagues [Eur. J. Immunol. 2013. 43: 889-896] in this issue of the European Journal of Immunology shows that AMPK, by sensing glucose stress, is an important upstream molecule of mammalian target of rapamycin (mTOR) complex 1 for memory CD8+ T-cell differentiation. This study provides new insights into how AMPK monitors energy stress to control effector and memory CD8+ T-cell formation as discussed in this Commentary.

CD8+ T cells are an important arm of adaptive immunity and control infections of intracellular bacteria and viruses by killing infected cells, as well as by producing cytokines [1, 2]. It is also apparent that CD8+ T cells play an essential role in regulating cancer [3, 4]. An understanding of how highly functional memory CD8+ T cells are formed from their naïve predecessors is one of the crucial factors required to develop a successful vaccine against chronic infections and cancer. During an acute infection, antigen-specific naïve CD8+ T cells activated by antigen-presenting cells rapidly proliferate and differentiate into effector cells 1–2 weeks post infection. After pathogen clearance, the majority of these effector cells die during the effector-to-memory transition phase, but a small fraction (5–10%) of the effector cells survive and differentiate into memory CD8+ T cells. These surviving effector cells are known to be memory precursor cells that can be distinguished by IL-7R and KLRG1 expression [5-7]. Memory CD8+ T cells formed from memory precursor cells further differentiate into self-renewing memory T cells with an enhanced recall capacity.

During this differentiation process, metabolic activity in CD8+ T cells dramatically changes [8, 9]. Although memory T cells are qualitatively distinct from naïve T cells in terms of cytokine production, longevity, and localization etc., both naïve and memory cells are metabolically quiescent, displaying low rates of amino acid/glucose uptake and protein synthesis [8, 9]. In contrast, proliferating effector T cells substantially enhance glucose uptake to support massive expansion and biosynthetic demands, and become metabolically active and produce effector molecules by switching the energy production mechanism [8, 9]; mitochondrial oxidative phosphorylation is primarily used to generate ATP in quiescent cells while energy production in activated cells is dependent on the glycolytic pathway. Since this metabolic reprogramming accompanies the differentiation of naïve T cells into effector/memory T cells, it seems that these metabolic changes have crucial roles in memory T-cell formation. Supporting this notion, recent studies have shown that mTOR complex 1 (mTORC1) regulates memory CD8+ T-cell differentiation [10, 11].

mTOR is a serine/threonine kinase that controls cell growth and proliferation, and forms two distinct complexes, mTORC1 and mTORC2. mTORC1 signals can enhance glycolysis by upregulating several transcription factors such as HIF1a and MYC [12, 13]. Inhibiting mTORC1 with rapamycin or RNAi in antigen-specific CD8+ T cells significantly enhances both the quantity and quality of memory CD8+ T cells [10]. Its inhibition in such cells during the T-cell expansion phase (naïve to effector) increases the formation of memory precursor effector cells that have an IL-7Rhi KLRG1lo phenotype while, during the T-cell contraction phase, the differentiation of effector to highly functional memory T cells with self-renewing ability was accelerated by inhibiting mTORC1 activity [10]. Following on from these findings, the role of T-cell metabolism in memory differentiation has received considerable attention.

Various upstream molecules (TSC1/2, Rheb, LKB1 etc.) that regulate mTORC1 have already been identified, but less is known about the biological role of such molecules in memory CD8+ T-cell formation [14]. Adenosine monophosophate-activated protein kinase (AMPK), an upstream inhibitor of mTORC1, is an evolutionally conserved kinase that regulates cellular energy homeostasis by inhibiting anabolic energy-consuming metabolism [15, 16]. Experiments using mice lacking AMPK have revealed that this kinase is dispensable for the activation of T cells stimulated with anti-CD3/CD28 antibodies in vitro as well as stimulation with peptides in vivo [17, 18]. In addition, another recent paper has examined by the use of metformin, a drug that activates AMPK, whether AMPK regulates memory CD8+ T-cell differentiation [11]. Administration of this drug into Listeria monocytogenes (LM) infected mice promoted memory CD8+ T-cell formation suggesting that AMPK plays an important role in memory CD8+ T-cell differentiation [11]. However, because AMPK is expressed in a number of tissues and metformin has many effects on cells that are unrelated to its actions on AMPK, it is not clear whether the enhanced memory T-cell formation seen in metformin-treated LM-infected mice is caused by the inhibition of AMPK in antigen-specific CD8+ T cells.

In this issue of the European Journal of Immunology, Cantrell and colleagues [19] tackle the above question and examine the role of AMPK in memory CD8+ T-cell differentiation. They used mice with floxed AMPK-α1 alleles, and deleted AMPK-α1 by backcrossing to transgenic mice expressing Cre recombinase under the control of the CD4 promoter. To examine antigen-specific CD8+ T-cell responses, AMPK-α1-deficient OT-1 cells were also generated using the same system. The authors found that AMPK-α1 was not required for activation of CD8+ T cells upon antigen stimulation. Thus, phenotypic changes, proliferation, and IFN-γ production of AMPK-α1null OT-1 cells stimulated with peptide in vitro were quite similar to those of wild-type OT-1 cells [19]. Furthermore, the authors confirmed that AMPK-α1null CD8+ T cells could differentiate into effector cells in vivo [19]. LM-OVA-infected mice into which AMPK-α1null OT-1 cells were adoptively transferred induced effector T cells to levels comparable to those induced by wild-type OT-1 cells [19]. These results differ from those of MacIver et al. [17] who showed that AMPK-α1 deficiency in CD8+ T cells promotes hyperactivation characterized by upregulated CD44 expression, as well as enhanced IFN-γ production; however, these discrepancies can be explained by the system used in the two independent studies: Cantrell and colleagues [19] conditionally knocked out AMPK-α1 in T cells using AMPK-α1f/f CD4 cre mice while MacIver et al. [17] examined AMPK-α1-deficient T cells obtained from AMPK-α1 knockout mice. Given that all cells in the knockout mice lack the AMPK-α1 gene, the development of T cells in such mice may be altered as compared with wild-type and AMPK-α1f/f CD4 Cre mice. Indeed, CD8+ T cells isolated from AMPK-α1 knockout mice show elevated levels of CD44 expression [17], suggesting that they have a higher number of antigen-experienced CD8+ T cells (e.g. effector cells) in the periphery as compared with wild-type mice. Thus, these T cells in AMPK-α1 knockout mice may lead to hyperactivation of CD8+ T cells after stimulation. Despite this discrepancy, both studies [17, 19] show that AMPK-α1 is dispensable for the activation of CD8+ T cells and suggest that its activity has minimal impact on the generation of effector CD8+ T cells.

Interestingly, Cantrell and colleagues [19] find that AMPK-α1 plays a critical role in memory CD8+ T-cell differentiation during the effector-to-memory transition phase. Effector T cells generated by antigenic stimulation and cytokines express high levels of the glucose transporter [8] and can efficiently uptake glucose for glycolysis metabolism. The current study [19] shows that AMPK-α1 in effector T cells acts as a metabolic switch to regulate mTORC1 activity by sensing glucose stress. The authors observed that mTORC1 was highly active in effector T cells in the presence of glucose, and that deprivation of glucose inhibited the activity [19]. Furthermore, AMPK-α1, an upstream inhibitor of mTORC1, was severely inhibited in medium with adequate amounts of glucose, and its activity was recovered by withdrawal of glucose [19]. Notably, AMPK-α1null effector T cells maintained high mTORC1 activity in the absence of glucose [19], indicating that AMPK-α1 in effector CD8+ T cells is required for termination of mTORC1 activity under conditions of glucose stress. This explains, at least in part, the reason why AMPK-α1 is dispensable for T-cell activation and proliferation. Activated T cells are able to uptake a high amount of glucose through increased expression of the glucose transporter, which in turn can activate mTORC1 by significantly inhibiting AMPK activity (Fig. 1). Thus, it seems that AMPK activity in T cells is consistently inhibited during the naïve-to-effector transition phase and has minimal impact on the generation of effector T cells (Fig. 1).


Figure 1. Proposed model for the regulation of memory CD8+ T-cell differentiation by AMPK. During the naïve-to-effector transition phase, antigenic stimulation and inflammatory cytokines upregulate the expression of the glucose transporter, and thereby enable CD8+ T cells to increase glucose uptake. The subsequent high levels of glucose indirectly inhibit AMPK, and this allows activation of mTORC1 to promote activation and proliferation, as well as effector differentiation, of T cells. On the other hand, during the effector-to-memory transition phase, expression levels of the glucose transporter decrease due to a lack of antigenic stimulation and minimal levels of inflammatory cytokines. The consequent reduced levels of glucose in the cytosol cause activation of AMPK, which leads to the inhibition of mTORC1 and a change in the metabolic state of the effector T cells to that of quiescence, and the promotion of memory CD8+ T-cell differentiation.

Download figure to PowerPoint

Since AMPK appears to regulate mTORC1 which in turn influences memory CD8+ T-cell differentiation and cellular metabolism, AMPK might have an important role in memory T-cell formation during the effector-to-memory T-cell transition. To address this issue, Cantrell and colleagues [19] generated wild-type and AMPK-α1null effector CD8+ T cells in vitro and co-transferred equal numbers of these effector T cells into naïve mice. In these recipient naive mice, the effector T cells downregulate the expression of the glucose and amino acid transporters due to a lack of antigenic stimulation and inflammatory cytokines (Fig. 1), and hence the availability of such nutrients in these T cells significantly decreases. Thus, the effector T cells need to switch their metabolism to that seen in quiescent cells to ensure their survival and hence subsequent memory T-cell formation. As better survival of the wild-type effector cells as compared with the AMPK-α1null effector cells was observed [19], it suggests that AMPK-α1null effector T cells lose their ability to return to a quiescent state. These results also imply that memory CD8+ T-cell differentiation of AMPK-α1null effector T cells might be altered due to the dysregulation of metabolic changes.

To examine this, the authors generated memory CD8+ T cells by LM infection [19]. Although IL-7R and KLRG1 expression levels were found to be similar between wild-type and AMPK-α1null memory CD8+ T cells, there was a striking difference in the recall responses of these cells. Experiments in which equal numbers of wild-type and AMPK-α1null OT-1 memory cells were adoptively co-transferred into naïve mice followed by LM-OVA challenge generated two to threefold higher numbers of wild-type OT-1, as compared with AMPK-α1null OT-1, memory cells at day 6 post infection [19]. These results are consistent with the previous observation that metformin, an activator of AMPK, enhances memory CD8+ T-cell differentiation [11]. Thus, AMPK seems to have an essential role in memory CD8+ T-cell differentiation, especially during the effector-to-memory transition phase (Fig. 1).

This study of Cantrell and colleagues [19] uncovers an important mechanism of AMPK, i.e. to regulate memory CD8+ T-cell formation, and indicates that molecules controlling mTOR activity significantly affect CD8+ T-cell responses. In addition to AMPK, there are various upstream molecules (TSC1/2, Rheb, LKB1 etc.) of mTOR [14], but the in vivo role of such individual molecules in memory CD8+ T-cell differentiation is largely unknown. Thus, it will be interesting to further study the upstream molecules and pathways of mTOR during memory CD8+ T-cell development for a better understanding of how mTOR-dependent metabolism is controlled in T cells.

Conflict of interest

  1. Top of page
  2. Abstract
  3. Conflict of interest
  4. References

The authors declare no financial or commercial conflict of interest.


  1. Top of page
  2. Abstract
  3. Conflict of interest
  4. References

adenosine monophosophate-activated protein kinase


Listeria monocytogenes


mTOR complex 1